A new opportunity for active managers

June 04, 2018

Structural and cyclical changes in industry benchmarks can mean these evolve into poor metrics for assessing active manager skill. We believe this is particularly a problem for Asian equities. The Asian MSCI benchmark has been morphing into an un-diversified momentum-biased portfolio. Nonetheless, emerging cycle changes could see a reversal in this dynamic. This has significant implications for ETF flows, style bias and active manager performance.


A new opportunity for active managers
Benchmarking is ubiquitous in the world of active fund management. The base reason for this is of course the need to attempt to discern what value the manager is adding over and above some comparative metric (commonly a relevant market index). Hence, the “free-ride” provided by the market return is deducted and the remainder (positive or negative) indicates manager skill. This process is logical, reasonable and expected within the industry. Unfortunately, it does not always work as intended.

Implicit in the selection of a benchmark for performance analysis is the assumption the benchmark is something an investor desires. The benchmark will likely reflect an appropriate asset class, an appropriate geographical region, and may also incorporate more specific characteristics (e.g. a small cap equity index, a style-biased index, etc) designed to increase the accuracy of manager assessment. However, the process can go awry.

An obvious limitation of benchmark setting is that it typically only occurs once in the lifetime of an investment product – at its initial formation. Changing the benchmark is operationally complex and may well raise red flags with existing and potential future investors given it may be construed as an attempt by the manager to take advantage of those investors. Yet changing the benchmark may be a completely rational decision, and in fact to the benefit of investors. Assuming the investment product is not changing, it may be appropriate to adjust the benchmark because the benchmark itself has undergone change – to the point where that benchmark no longer represents a desired outcome for investors.

For illustrative purposes, imagine an equity fund manager benchmarked against a region- and style-appropriate equity index (e.g. a MSCI equity index). The key is that the composition of the benchmark itself is not static. Benchmark composition may change passively as a result of relative performance of benchmark composites and may change actively due to adjustments made by the index guardians. MSCI, for example, frequently adds and removes stocks from its indices, is now beginning the process of adding China A shares to various benchmarks, and is soon to introduce substantial changes to sector definitions/composition.1

Figure 1 provides one illustration of the kinds of distortions generated by MSCI re-balancing. In this instance a company with a market cap in excess of USD3.5b surged 43% into the month-end market close, before dropping straight back down the following day, on the back of index inclusion.

These changes may be completely reasonable and have little to no impact on the relevance of the benchmark for measuring active manager skill. However, that is not always the case. Say we launched a poll of investors with medium- to long-term investment horizons, asking whether they desired either:

1) A momentum-chasing low-diversification portfolio; or,

2) A fundamentally-driven highly diversified portfolio.

We suspect sophisticated investors would likely be biased towards option (2), while there may be somewhat higher relative demand for option (1) amongst less sophisticated investors. Essentially, more sophisticated investors would appreciate the value of fundamental analysis and comprehensive risk management. The problem with this in Asia is that the benchmark regional MSCI index is option (1).

Some key illustrative metrics:

1) We estimate the MSCI Asia ex Japan index is over 30% Information Technology and two thirds of that is just five stocks, so five stocks are more than 20% of the entire regional index.

2) Those same five stocks have comprised around 50% of total regional index return over the past 1-2 years (and substantially more for lengthy periods within that).

3) The effect is more dramatic within countries. IT represents over 40% of the MSCI China index and nearly 90% of that is just three stocks.

ETF flows have aggravated this development, resulting in momentum-chasing. Benchmark ETFs have to buy more of stocks that are recording the strongest relative performance, driving further upside in those same stocks. This dynamic also contributed to a reduction in market volatility (albeit there were other factors involved in that as well, such as the “bread-and-butter” hedge fund trade of selling volatility). As a result, investors look like heroes twice over for chasing such performance – firstly in terms of return outcome and secondly in terms of risk.

However, the use of the word “risk” in this case is something of a misnomer. Investors look like they are generating a great risk outcome on the basis of backward-looking volatility. Unfortunately, particularly in places like Asia with plenty of emerging market exposure, such a measure of risk can be very poor. Historic volatility has only limited information for future volatility. So, investors are incentivized to chase passive index-benchmarked ETFs for both return and risk reasons, yet we know ultimately that theme has a natural limit.

In fact, we believe that limit has already been reached, although the unwinding will gradually materialize over a lengthy period. Key to this unwinding is the yield curve. Rising long bond yields, in particular, begin to bias stock selection away from momentum-chasing and more towards fundamental value. Rising short yields, reducing liquidity, add to the effect. This creates myriad opportunities for active managers, while passive ETFs suffer as performance begins to shift away from the momentum plays. Indeed, we may well see a significant shift away from ETFs and towards active managers over the next 1-2 years, reversing what has been an exceptionally strong multi-year trend. This would likely also encompass a shift away from growth-style investing towards value.

What would be a better way to benchmark? Ideally, benchmarks would be tailored to individual investor preferences and dynamics. These should match an investor’s desired risk characteristics (whether those be with reference to volatility, tail risk, expected shortfall, etc) and investment horizon. From those dynamics it would be possible to impute the desired return characteristics. Unfortunately, this can be a relative complex process. A sophisticated investor may be able to achieve these outcomes through an asset allocation process, but few investors would go to the extent of tailoring a process for individual holdings. However, by not doing so, the assessment process for manager skill may suffer.

So where does this leave us? All of the above implies that investors need to obtain not only a strong understanding of the strategy underlying an active manager’s process, but also recognize when the industry standard of simple index benchmarking may be a poor basis for assessment. In particular, simple index benchmarking can be an extremely poor assessment measure at cycle turning points. This is when the benchmark itself can be most mis-leading with regard to FUTURE risk and return. We may well be at that point now in Asia.

Of course we recognize that all of the above may simply be construed as a case of sour grapes. We manage an Asian systematic equity strategy benchmarked against the Asia ex Japan MSCI index. In the interests of full disclosure, we have been heavily underweight the mega-cap tech stocks which have driven most of the aggregate market performance over the last two years. Having said that, our strategy has delivered a net return close to the benchmark return, and done so with lower volatility. This is because the strategy has been specifically designed to:

1) Generate a return around benchmark, through its stock selection, when the style backdrop is against it (the strategy has been steadily shifting towards a value bias and lower beta as the mega-cap tech stocks surge higher);

2) Generate a premium return when the style backdrop is in its favour (i.e. adding value on stock selection and style bias); and,

3) Do (1) and (2) with lower-than-market volatility by maintaining a higher level of portfolio diversification than is evident in the benchmark.

Hence, we are very happy that our systematic strategy, now with more than two years of live trading, has performed in a manner entirely consistent with our historical testing.

Looking forward, if we continue to see US short and long rates head higher, we expect the active/passive dynamic discussed above to develop, along with a shift away from growth towards value. This should generate significant relative outperformance in the strategy albeit, as we have argued, relative to a flawed index.

Overall, this means investors need to pay special attention to whether or not the benchmark they are using to assess manager skill is something that, in its own right, represents a desired outcome. If that index reflects a concentrated momentum-biased portfolio, this may well not match with the requirements of a sophisticated institutional investor. On the plus side, even though the Asian MSCI benchmark may have flaws at present, changing cycle dynamics could see a significant reversal in those features. That in turn should be accompanied by a swing back towards active managers, away from ETFs, and towards the value style.

1 In the context of Asian (and emerging market) benchmarking the addition of China A shares to MSCI indices creates a range of challenges for fund managers. Although the additions at this stage are small in capitalization terms, they are only likely to increase over time. We currently include a selection of China A shares in our systematic strategy. We are also investigating optimal methods for incorporating these stocks in portfolio holdings. However, we also recognize there are huge problems with data for Chinese stocks. We include a range of checks for earnings manipulation by management of listed firms. These checks are akin to the analysis performed by activist short sellers. Offending stocks are removed from our investment universe. However, relative to other countries, China has a larger proportion of stocks raising red flags in our manipulation checks. We view the addition by MSCI of more China stocks to benchmark indices as troubling when forensic accounting highlights a significant proportion of these have serious data anomalies. Offenders include large firms.

Written by: Dr. Hamish Macalister